René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Functionality to fetch output schemas as well on list_tools()","articleBody":"**Is your feature request related to a problem? Please describe.**\nI was working on a project and was wondering if it was possible to fetch the output schemas as well on `client.list_tools()`. I realize that the output returned by MCP tools is often a string (or stringified JSON), and having information in advance about the structure of data I would be receiving would be helpful. \n\n**Describe the solution you'd like**\nCurrently the `list_tools()` returns something like :\n```\n {\n        \"meta\": null,\n        \"nextCursor\": null,\n        \"tools\": [\n          {\n            \"name\": \"get-news\",\n            \"description\": \"Get the latest news for a given topic\",\n            \"input_schema\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"topic\": {\n                  \"type\": \"string\",\n                  \"description\": \"Topic to search for\"\n                },\n                \"page_size\": {\n                  \"type\": \"number\",\n                  \"description\": \"Number of results to return (1-25)\"\n                }\n              },\n              \"required\": [\n                \"topic\",\n                \"page_size\"\n              ]\n            },\n            \"annotations\": null\n          },\n          {\n            \"name\": \"get-headlines\",\n            \"description\": \"Get the latest headlines for a given country\",\n            \"input_schema\": {\n              \"type\": \"object\",\n              \"properties\": {\n                \"country\": {\n                  \"type\": \"string\",\n                  \"description\": \"Country code (2 letters)\"\n                },\n                \"page_size\": {\n                  \"type\": \"number\",\n                  \"description\": \"Number of results to return (1-25)\"\n                }\n              },\n              \"required\": [\n                \"country\",\n                \"page_size\"\n              ]\n            },\n            \"annotations\": null\n          }\n        ]\n      }\n\n```\nI'd like to also get the output schema for each tool alongside the input schema. If the output is a simple string, then just `{\"type\": \"string\"}` would be fine. But for JSON objects, getting the actual JSON schema being returned before stringification would be very helpful.\n\n**Describe alternatives you've considered**\nCurrent alternatives involve manually reading the MCP server code to determine the exact output schema (this is repetitive for each new MCP server we add) or sending a test request and then parsing the output to infer the schema (not feasible in many cases).\n\nSince I'm using MCP with the client only, I don't have an agent to interpret the output on runtime and utilize that information dynamically.","author":{"url":"https://github.com/PratyushNag","@type":"Person","name":"PratyushNag"},"datePublished":"2025-05-19T13:52:24.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":2},"url":"https://github.com/754/python-sdk/issues/754"}